USB: EHCI: split ehci_qh into hw and sw parts
The ehci_qh structure merged hw and sw together which is not good:
1. More and more items are being added into ehci_qh, the ehci_qh software
part are unnecessary to be allocated in DMA qh_pool.
2. If HCD has local SRAM, the sw part will consume it too, and it won't
bring any benefit.
3. For non-cache-coherence system, the entire ehci_qh is uncachable, actually
we only need the hw part to be uncacheable. Spliting them will let the sw
part to be cacheable.
Signed-off-by: Alek Du <alek.du@intel.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
CC: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
diff --git a/drivers/usb/host/ehci-sched.c b/drivers/usb/host/ehci-sched.c
index edd61ee..327437a 100644
--- a/drivers/usb/host/ehci-sched.c
+++ b/drivers/usb/host/ehci-sched.c
@@ -60,6 +60,20 @@
}
}
+static __hc32 *
+shadow_next_periodic(struct ehci_hcd *ehci, union ehci_shadow *periodic,
+ __hc32 tag)
+{
+ switch (hc32_to_cpu(ehci, tag)) {
+ /* our ehci_shadow.qh is actually software part */
+ case Q_TYPE_QH:
+ return &periodic->qh->hw->hw_next;
+ /* others are hw parts */
+ default:
+ return periodic->hw_next;
+ }
+}
+
/* caller must hold ehci->lock */
static void periodic_unlink (struct ehci_hcd *ehci, unsigned frame, void *ptr)
{
@@ -71,7 +85,8 @@
while (here.ptr && here.ptr != ptr) {
prev_p = periodic_next_shadow(ehci, prev_p,
Q_NEXT_TYPE(ehci, *hw_p));
- hw_p = here.hw_next;
+ hw_p = shadow_next_periodic(ehci, &here,
+ Q_NEXT_TYPE(ehci, *hw_p));
here = *prev_p;
}
/* an interrupt entry (at list end) could have been shared */
@@ -83,7 +98,7 @@
*/
*prev_p = *periodic_next_shadow(ehci, &here,
Q_NEXT_TYPE(ehci, *hw_p));
- *hw_p = *here.hw_next;
+ *hw_p = *shadow_next_periodic(ehci, &here, Q_NEXT_TYPE(ehci, *hw_p));
}
/* how many of the uframe's 125 usecs are allocated? */
@@ -93,18 +108,20 @@
__hc32 *hw_p = &ehci->periodic [frame];
union ehci_shadow *q = &ehci->pshadow [frame];
unsigned usecs = 0;
+ struct ehci_qh_hw *hw;
while (q->ptr) {
switch (hc32_to_cpu(ehci, Q_NEXT_TYPE(ehci, *hw_p))) {
case Q_TYPE_QH:
+ hw = q->qh->hw;
/* is it in the S-mask? */
- if (q->qh->hw_info2 & cpu_to_hc32(ehci, 1 << uframe))
+ if (hw->hw_info2 & cpu_to_hc32(ehci, 1 << uframe))
usecs += q->qh->usecs;
/* ... or C-mask? */
- if (q->qh->hw_info2 & cpu_to_hc32(ehci,
+ if (hw->hw_info2 & cpu_to_hc32(ehci,
1 << (8 + uframe)))
usecs += q->qh->c_usecs;
- hw_p = &q->qh->hw_next;
+ hw_p = &hw->hw_next;
q = &q->qh->qh_next;
break;
// case Q_TYPE_FSTN:
@@ -237,10 +254,10 @@
continue;
case Q_TYPE_QH:
if (same_tt(dev, q->qh->dev)) {
- uf = tt_start_uframe(ehci, q->qh->hw_info2);
+ uf = tt_start_uframe(ehci, q->qh->hw->hw_info2);
tt_usecs[uf] += q->qh->tt_usecs;
}
- hw_p = &q->qh->hw_next;
+ hw_p = &q->qh->hw->hw_next;
q = &q->qh->qh_next;
continue;
case Q_TYPE_SITD:
@@ -375,6 +392,7 @@
for (; frame < ehci->periodic_size; frame += period) {
union ehci_shadow here;
__hc32 type;
+ struct ehci_qh_hw *hw;
here = ehci->pshadow [frame];
type = Q_NEXT_TYPE(ehci, ehci->periodic [frame]);
@@ -385,17 +403,18 @@
here = here.itd->itd_next;
continue;
case Q_TYPE_QH:
+ hw = here.qh->hw;
if (same_tt (dev, here.qh->dev)) {
u32 mask;
mask = hc32_to_cpu(ehci,
- here.qh->hw_info2);
+ hw->hw_info2);
/* "knows" no gap is needed */
mask |= mask >> 8;
if (mask & uf_mask)
break;
}
- type = Q_NEXT_TYPE(ehci, here.qh->hw_next);
+ type = Q_NEXT_TYPE(ehci, hw->hw_next);
here = here.qh->qh_next;
continue;
case Q_TYPE_SITD:
@@ -498,7 +517,8 @@
dev_dbg (&qh->dev->dev,
"link qh%d-%04x/%p start %d [%d/%d us]\n",
- period, hc32_to_cpup(ehci, &qh->hw_info2) & (QH_CMASK | QH_SMASK),
+ period, hc32_to_cpup(ehci, &qh->hw->hw_info2)
+ & (QH_CMASK | QH_SMASK),
qh, qh->start, qh->usecs, qh->c_usecs);
/* high bandwidth, or otherwise every microframe */
@@ -517,7 +537,7 @@
if (type == cpu_to_hc32(ehci, Q_TYPE_QH))
break;
prev = periodic_next_shadow(ehci, prev, type);
- hw_p = &here.qh->hw_next;
+ hw_p = shadow_next_periodic(ehci, &here, type);
here = *prev;
}
@@ -528,14 +548,14 @@
if (qh->period > here.qh->period)
break;
prev = &here.qh->qh_next;
- hw_p = &here.qh->hw_next;
+ hw_p = &here.qh->hw->hw_next;
here = *prev;
}
/* link in this qh, unless some earlier pass did that */
if (qh != here.qh) {
qh->qh_next = here;
if (here.qh)
- qh->hw_next = *hw_p;
+ qh->hw->hw_next = *hw_p;
wmb ();
prev->qh = qh;
*hw_p = QH_NEXT (ehci, qh->qh_dma);
@@ -581,7 +601,7 @@
dev_dbg (&qh->dev->dev,
"unlink qh%d-%04x/%p start %d [%d/%d us]\n",
qh->period,
- hc32_to_cpup(ehci, &qh->hw_info2) & (QH_CMASK | QH_SMASK),
+ hc32_to_cpup(ehci, &qh->hw->hw_info2) & (QH_CMASK | QH_SMASK),
qh, qh->start, qh->usecs, qh->c_usecs);
/* qh->qh_next still "live" to HC */
@@ -596,6 +616,7 @@
static void intr_deschedule (struct ehci_hcd *ehci, struct ehci_qh *qh)
{
unsigned wait;
+ struct ehci_qh_hw *hw = qh->hw;
qh_unlink_periodic (ehci, qh);
@@ -606,14 +627,14 @@
*/
if (list_empty (&qh->qtd_list)
|| (cpu_to_hc32(ehci, QH_CMASK)
- & qh->hw_info2) != 0)
+ & hw->hw_info2) != 0)
wait = 2;
else
wait = 55; /* worst case: 3 * 1024 */
udelay (wait);
qh->qh_state = QH_STATE_IDLE;
- qh->hw_next = EHCI_LIST_END(ehci);
+ hw->hw_next = EHCI_LIST_END(ehci);
wmb ();
}
@@ -739,14 +760,15 @@
unsigned uframe;
__hc32 c_mask;
unsigned frame; /* 0..(qh->period - 1), or NO_FRAME */
+ struct ehci_qh_hw *hw = qh->hw;
qh_refresh(ehci, qh);
- qh->hw_next = EHCI_LIST_END(ehci);
+ hw->hw_next = EHCI_LIST_END(ehci);
frame = qh->start;
/* reuse the previous schedule slots, if we can */
if (frame < qh->period) {
- uframe = ffs(hc32_to_cpup(ehci, &qh->hw_info2) & QH_SMASK);
+ uframe = ffs(hc32_to_cpup(ehci, &hw->hw_info2) & QH_SMASK);
status = check_intr_schedule (ehci, frame, --uframe,
qh, &c_mask);
} else {
@@ -784,11 +806,11 @@
qh->start = frame;
/* reset S-frame and (maybe) C-frame masks */
- qh->hw_info2 &= cpu_to_hc32(ehci, ~(QH_CMASK | QH_SMASK));
- qh->hw_info2 |= qh->period
+ hw->hw_info2 &= cpu_to_hc32(ehci, ~(QH_CMASK | QH_SMASK));
+ hw->hw_info2 |= qh->period
? cpu_to_hc32(ehci, 1 << uframe)
: cpu_to_hc32(ehci, QH_SMASK);
- qh->hw_info2 |= c_mask;
+ hw->hw_info2 |= c_mask;
} else
ehci_dbg (ehci, "reused qh %p schedule\n", qh);
@@ -2188,7 +2210,7 @@
case Q_TYPE_QH:
/* handle any completions */
temp.qh = qh_get (q.qh);
- type = Q_NEXT_TYPE(ehci, q.qh->hw_next);
+ type = Q_NEXT_TYPE(ehci, q.qh->hw->hw_next);
q = q.qh->qh_next;
modified = qh_completions (ehci, temp.qh);
if (unlikely (list_empty (&temp.qh->qtd_list)))